Cerebras Inference Chat Completions API
CloseCerebras Inference Chat Completions API
Cerebras 官方参考页定义聊天生成 POST API,需 Bearer API key,输入 model 与 messages,返回 choices、usage 和计时信息;支持流式及工具调用但参数受模型限制。费用、配额和数据保留未知,仅文档核对。
Source statements below remain in their published language. Collection, contributor reports and a provider's identity are separate; this page does not certify quality, current availability or permission to execute.
- Stable ID
- cap-67a2f1e31211aa65422c0b60
- Card version
- 2
- Updated
- 2026-10-24T17:49:02.2465278+00:00
- Ownership state
- unclaimed
- Processing
- unknown
- Evidence level
- unknown
- Business execution
- unknown
Keep this fixed version · Original public source · Complete machine record
Published description
该接口面向聊天助手和多轮对话,按消息角色提交上下文并生成 completion。正文给出字段与响应参考,附件内嵌的官方 bash 示例明确完整 URL。本次只整理文档,不获取 key、安装 SDK、调用推理或执行返回的工具请求。
Inputs
- 必填 model 字符串和 messages 数组;role 支持 system、user、assistant、tool、developer,文本或图像输入支持依模型。
- 可选 stream、temperature、top_p、max_completion_tokens、reasoning_effort、response_format、tools 与 tool_choice 等。
- Content-Type 支持 application/json 或 application/vnd.msgpack;可选 gzip Content-Encoding。
Outputs
- 成功 200 响应包含 id、choices、created、model、object,以及 usage 和 time_info。
- choices.message 可含 content、reasoning 或 tool_calls;finish_reason 包括 stop、length、content_filter、tool_calls。
- stream=true 返回局部 message deltas;tool_calls 是模型请求,本次没有执行其所描述的函数。
Invocation and authentication
{
"authentication": {
"kind": "api_key",
"notes": "\u53C2\u8003\u9875\u8981\u6C42 API key\uFF0C\u4ECE Cerebras Cloud console \u83B7\u53D6\uFF0C\u5E76\u4EE5 Bearer header \u4F20\u5165\uFF1B\u672C\u6B21\u672A\u53D6\u5F97\u6216\u4F7F\u7528\u51ED\u636E\u3002"
},
"command": null,
"entry_url": "https://api.cerebras.ai/v1/chat/completions",
"kind": "http_api",
"method": "POST",
"request_example": {
"body": {
"messages": [
{
"content": "Hello!",
"role": "user"
}
],
"model": "qwen-3.8-27b",
"reasoning_effort": "none"
},
"method": "POST",
"notes": "\u5B8C\u6574 URL \u6765\u81EA\u672C\u9644\u4EF6\u5185\u5D4C\u5B98\u65B9 bash \u793A\u4F8B\uFF1B\u53E6\u9700\u771F\u5B9E Bearer \u51ED\u636E\u3002\u4EC5\u793A\u4F8B\uFF0C\u672A\u8C03\u7528\u3002",
"url": "https://api.cerebras.ai/v1/chat/completions"
},
"requirements": [
"\u9700\u6709\u6548\u7684 Cerebras Cloud API key \u4E0E\u53EF\u7528 model ID\u3002",
"JSON body \u7684 model\u3001messages \u4E3A\u5FC5\u586B\uFF1B\u6D88\u606F\u5185\u5BB9\u4E0E\u53C2\u6570\u652F\u6301\u53D6\u51B3\u4E8E\u6240\u9009\u6A21\u578B\u3002",
"\u53EF\u76F4\u63A5\u4F7F\u7528 HTTP\uFF1BPython / Node SDK \u53EA\u662F\u6587\u6863\u793A\u4F8B\u8DEF\u5F84\uFF0C\u672C\u6B21\u672A\u5B89\u88C5\u3002"
]
}Commands and request examples are documentation only; this page does not execute them. Confirm external installation, disclosure and fees separately.
Third-party pricing and data handling
{
"model": "unknown",
"notes": "\u6240\u6355\u83B7\u7684 Chat Completions \u53C2\u8003\u6CA1\u6709\u8D39\u7387\u6216\u514D\u8D39\u989D\u5EA6\uFF1Busage \u4E3A\u7528\u91CF\u5B57\u6BB5\uFF0C\u4E0D\u80FD\u7531\u5176\u6837\u4F8B\u63A8\u65AD\u5B9E\u9645\u4EF7\u683C\u3002",
"source_url": "https://inference-docs.cerebras.ai/api-reference/chat-completions"
}{
"location": "provider_server",
"notes": "\u6D88\u606F\u4E0E\u53EF\u9009\u56FE\u50CF\u7531 Cerebras \u670D\u52A1\u7AEF\u5904\u7406\u3002\u6587\u6863\u63D0\u5230 prompt caching \u4F1A\u81EA\u52A8\u5DE5\u4F5C\uFF0C\u4F46\u672C\u9875\u6CA1\u6709\u4FDD\u7559\u671F\u9650\u3001\u8BAD\u7EC3\u4F7F\u7528\u6216\u9690\u79C1\u627F\u8BFA\uFF1B\u672C\u6B21\u672A\u63D0\u4EA4\u5185\u5BB9\u3002"
}Unknown is not a claim that a service is free, anonymous, private or unavailable.
Limitations
- 当前 n 仅支持1;max_tokens 与 max_completion_tokens 是别名,不能同时发送。
- 模型对 reasoning、图像、工具、输出格式及 service_tier 的支持有差异,不能把单模型样例推广到全部模型。
- json_object 不兼容 streaming;seed 仅尽力稳定,不保证确定性。
- 文档描述 service tier 和 quota 概念,但未给此账号的数值限额、价格或 SLA。
Unknowns
- 真实 key 和账号/model 权限、数值限流、配额、价格及 SLA 未验证。
- 消息/图像/缓存的保留期限、训练使用和隐私政策不在本附件。
- 样例延迟与 token 数是文档样例,不是本次测试或性能保证。
Fixed profiles, conditions and evidence
Read state: not_checked; review state: unknown. Do not combine different profiles.
资料与来源声明不等于本站实测;未列出的条件仍未知。
结构化配方只按公开声明检查约束与接口,不执行任务或授予权限;缺失记录保持未知,接口兼容不证明实际成功。
Contributor-reported verification
{
"checked_at": "2026-10-24T17:44:44.8693014\u002B00:00",
"checks_performed": [
"\u6765\u6E90HTTPS\u54CD\u5E94\u53EF\u8BFB\u53D6"
],
"evidence_urls": [
"https://inference-docs.cerebras.ai/api-reference/chat-completions"
],
"level": "documentation_only",
"not_verified": [
"\u529F\u80FD\u3001\u9274\u6743\u3001\u5B89\u88C5\u3001\u8F93\u51FA\u8D28\u91CF\u3001\u957F\u671F\u53EF\u7528\u6027\u5747\u672A\u7531\u672C\u7AD9\u6D4B\u8BD5\u3002"
],
"outcome": "not_tested",
"reported_by": "contributor",
"software_version": null,
"test_summary": "\u4EC5\u8BFB\u53D6\u516C\u5F00\u6765\u6E90\uFF1B\u672A\u8C03\u7528\u6240\u8FF0\u5DE5\u5177\u3002"
}Connection, installation, execution, output checks and user acceptance are different stages. This is not a commercial call or income receipt.
Dynamic observations and corrections
null
null
Add a public correction or observation
Submit only facts you may disclose, without private tasks, customer data or credentials. Corrections are proposals; observations are conditional contributor reports, not automatically verified facts, commercial calls or income.
Propose a sourced summary correction
Report a conditional public observation
Complete fixed public record and original source fields
This is the existing public read projection, not a translated or rewritten snapshot. Dynamic observations, where present, remain labelled separately from evidence.
{
"author_name": "Cerebras",
"availability": "unknown",
"callability": "documentation_only",
"category": "development",
"data_handling": {
"location": "provider_server",
"notes": "\u6D88\u606F\u4E0E\u53EF\u9009\u56FE\u50CF\u7531 Cerebras \u670D\u52A1\u7AEF\u5904\u7406\u3002\u6587\u6863\u63D0\u5230 prompt caching \u4F1A\u81EA\u52A8\u5DE5\u4F5C\uFF0C\u4F46\u672C\u9875\u6CA1\u6709\u4FDD\u7559\u671F\u9650\u3001\u8BAD\u7EC3\u4F7F\u7528\u6216\u9690\u79C1\u627F\u8BFA\uFF1B\u672C\u6B21\u672A\u63D0\u4EA4\u5185\u5BB9\u3002"
},
"description": "\u8BE5\u63A5\u53E3\u9762\u5411\u804A\u5929\u52A9\u624B\u548C\u591A\u8F6E\u5BF9\u8BDD\uFF0C\u6309\u6D88\u606F\u89D2\u8272\u63D0\u4EA4\u4E0A\u4E0B\u6587\u5E76\u751F\u6210 completion\u3002\u6B63\u6587\u7ED9\u51FA\u5B57\u6BB5\u4E0E\u54CD\u5E94\u53C2\u8003\uFF0C\u9644\u4EF6\u5185\u5D4C\u7684\u5B98\u65B9 bash \u793A\u4F8B\u660E\u786E\u5B8C\u6574 URL\u3002\u672C\u6B21\u53EA\u6574\u7406\u6587\u6863\uFF0C\u4E0D\u83B7\u53D6 key\u3001\u5B89\u88C5 SDK\u3001\u8C03\u7528\u63A8\u7406\u6216\u6267\u884C\u8FD4\u56DE\u7684\u5DE5\u5177\u8BF7\u6C42\u3002",
"inputs": [
"\u5FC5\u586B model \u5B57\u7B26\u4E32\u548C messages \u6570\u7EC4\uFF1Brole \u652F\u6301 system\u3001user\u3001assistant\u3001tool\u3001developer\uFF0C\u6587\u672C\u6216\u56FE\u50CF\u8F93\u5165\u652F\u6301\u4F9D\u6A21\u578B\u3002",
"\u53EF\u9009 stream\u3001temperature\u3001top_p\u3001max_completion_tokens\u3001reasoning_effort\u3001response_format\u3001tools \u4E0E tool_choice \u7B49\u3002",
"Content-Type \u652F\u6301 application/json \u6216 application/vnd.msgpack\uFF1B\u53EF\u9009 gzip Content-Encoding\u3002"
],
"invocation": {
"authentication": {
"kind": "api_key",
"notes": "\u53C2\u8003\u9875\u8981\u6C42 API key\uFF0C\u4ECE Cerebras Cloud console \u83B7\u53D6\uFF0C\u5E76\u4EE5 Bearer header \u4F20\u5165\uFF1B\u672C\u6B21\u672A\u53D6\u5F97\u6216\u4F7F\u7528\u51ED\u636E\u3002"
},
"command": null,
"entry_url": "https://api.cerebras.ai/v1/chat/completions",
"kind": "http_api",
"method": "POST",
"request_example": {
"body": {
"messages": [
{
"content": "Hello!",
"role": "user"
}
],
"model": "qwen-3.8-27b",
"reasoning_effort": "none"
},
"method": "POST",
"notes": "\u5B8C\u6574 URL \u6765\u81EA\u672C\u9644\u4EF6\u5185\u5D4C\u5B98\u65B9 bash \u793A\u4F8B\uFF1B\u53E6\u9700\u771F\u5B9E Bearer \u51ED\u636E\u3002\u4EC5\u793A\u4F8B\uFF0C\u672A\u8C03\u7528\u3002",
"url": "https://api.cerebras.ai/v1/chat/completions"
},
"requirements": [
"\u9700\u6709\u6548\u7684 Cerebras Cloud API key \u4E0E\u53EF\u7528 model ID\u3002",
"JSON body \u7684 model\u3001messages \u4E3A\u5FC5\u586B\uFF1B\u6D88\u606F\u5185\u5BB9\u4E0E\u53C2\u6570\u652F\u6301\u53D6\u51B3\u4E8E\u6240\u9009\u6A21\u578B\u3002",
"\u53EF\u76F4\u63A5\u4F7F\u7528 HTTP\uFF1BPython / Node SDK \u53EA\u662F\u6587\u6863\u793A\u4F8B\u8DEF\u5F84\uFF0C\u672C\u6B21\u672A\u5B89\u88C5\u3002"
]
},
"limitations": [
"\u5F53\u524D n \u4EC5\u652F\u63011\uFF1Bmax_tokens \u4E0E max_completion_tokens \u662F\u522B\u540D\uFF0C\u4E0D\u80FD\u540C\u65F6\u53D1\u9001\u3002",
"\u6A21\u578B\u5BF9 reasoning\u3001\u56FE\u50CF\u3001\u5DE5\u5177\u3001\u8F93\u51FA\u683C\u5F0F\u53CA service_tier \u7684\u652F\u6301\u6709\u5DEE\u5F02\uFF0C\u4E0D\u80FD\u628A\u5355\u6A21\u578B\u6837\u4F8B\u63A8\u5E7F\u5230\u5168\u90E8\u6A21\u578B\u3002",
"json_object \u4E0D\u517C\u5BB9 streaming\uFF1Bseed \u4EC5\u5C3D\u529B\u7A33\u5B9A\uFF0C\u4E0D\u4FDD\u8BC1\u786E\u5B9A\u6027\u3002",
"\u6587\u6863\u63CF\u8FF0 service tier \u548C quota \u6982\u5FF5\uFF0C\u4F46\u672A\u7ED9\u6B64\u8D26\u53F7\u7684\u6570\u503C\u9650\u989D\u3001\u4EF7\u683C\u6216 SLA\u3002"
],
"outputs": [
"\u6210\u529F 200 \u54CD\u5E94\u5305\u542B id\u3001choices\u3001created\u3001model\u3001object\uFF0C\u4EE5\u53CA usage \u548C time_info\u3002",
"choices.message \u53EF\u542B content\u3001reasoning \u6216 tool_calls\uFF1Bfinish_reason \u5305\u62EC stop\u3001length\u3001content_filter\u3001tool_calls\u3002",
"stream=true \u8FD4\u56DE\u5C40\u90E8 message deltas\uFF1Btool_calls \u662F\u6A21\u578B\u8BF7\u6C42\uFF0C\u672C\u6B21\u6CA1\u6709\u6267\u884C\u5176\u6240\u63CF\u8FF0\u7684\u51FD\u6570\u3002"
],
"pricing": {
"model": "unknown",
"notes": "\u6240\u6355\u83B7\u7684 Chat Completions \u53C2\u8003\u6CA1\u6709\u8D39\u7387\u6216\u514D\u8D39\u989D\u5EA6\uFF1Busage \u4E3A\u7528\u91CF\u5B57\u6BB5\uFF0C\u4E0D\u80FD\u7531\u5176\u6837\u4F8B\u63A8\u65AD\u5B9E\u9645\u4EF7\u683C\u3002",
"source_url": "https://inference-docs.cerebras.ai/api-reference/chat-completions"
},
"provider_claimed": false,
"record_kind": "public_source",
"representative_queries": [],
"sources": [
{
"basis": "official_documentation",
"checked_at": "2026-10-25T01:44:44.8693014\u002B08:00",
"title": "Chat Completions - Cerebras Inference",
"url": "https://inference-docs.cerebras.ai/api-reference/chat-completions"
}
],
"summary": "Cerebras \u5B98\u65B9\u53C2\u8003\u9875\u5B9A\u4E49\u804A\u5929\u751F\u6210 POST API\uFF0C\u9700 Bearer API key\uFF0C\u8F93\u5165 model \u4E0E messages\uFF0C\u8FD4\u56DE choices\u3001usage \u548C\u8BA1\u65F6\u4FE1\u606F\uFF1B\u652F\u6301\u6D41\u5F0F\u53CA\u5DE5\u5177\u8C03\u7528\u4F46\u53C2\u6570\u53D7\u6A21\u578B\u9650\u5236\u3002\u8D39\u7528\u3001\u914D\u989D\u548C\u6570\u636E\u4FDD\u7559\u672A\u77E5\uFF0C\u4EC5\u6587\u6863\u6838\u5BF9\u3002",
"tags": [
"http_api",
"documentation_only",
"llm_inference"
],
"terms_url": null,
"title": "Cerebras Inference Chat Completions API",
"unknowns": [
"\u771F\u5B9E key \u548C\u8D26\u53F7/model \u6743\u9650\u3001\u6570\u503C\u9650\u6D41\u3001\u914D\u989D\u3001\u4EF7\u683C\u53CA SLA \u672A\u9A8C\u8BC1\u3002",
"\u6D88\u606F/\u56FE\u50CF/\u7F13\u5B58\u7684\u4FDD\u7559\u671F\u9650\u3001\u8BAD\u7EC3\u4F7F\u7528\u548C\u9690\u79C1\u653F\u7B56\u4E0D\u5728\u672C\u9644\u4EF6\u3002",
"\u6837\u4F8B\u5EF6\u8FDF\u4E0E token \u6570\u662F\u6587\u6863\u6837\u4F8B\uFF0C\u4E0D\u662F\u672C\u6B21\u6D4B\u8BD5\u6216\u6027\u80FD\u4FDD\u8BC1\u3002"
],
"verification": {
"checked_at": "2026-10-24T17:44:44.8693014\u002B00:00",
"checks_performed": [
"\u6765\u6E90HTTPS\u54CD\u5E94\u53EF\u8BFB\u53D6"
],
"evidence_urls": [
"https://inference-docs.cerebras.ai/api-reference/chat-completions"
],
"level": "documentation_only",
"not_verified": [
"\u529F\u80FD\u3001\u9274\u6743\u3001\u5B89\u88C5\u3001\u8F93\u51FA\u8D28\u91CF\u3001\u957F\u671F\u53EF\u7528\u6027\u5747\u672A\u7531\u672C\u7AD9\u6D4B\u8BD5\u3002"
],
"outcome": "not_tested",
"reported_by": "contributor",
"software_version": null,
"test_summary": "\u4EC5\u8BFB\u53D6\u516C\u5F00\u6765\u6E90\uFF1B\u672A\u8C03\u7528\u6240\u8FF0\u5DE5\u5177\u3002"
},
"schema_version": "navonavo.capability.v1",
"service_id": "cap-67a2f1e31211aa65422c0b60",
"version": 2,
"source_url": "https://inference-docs.cerebras.ai/api-reference/chat-completions",
"ownership_state": "unclaimed",
"submitted_by": {
"kind": "importer",
"name": "NAVONAVO public material importer"
},
"submitter_identity_verified": false,
"source_check": {
"url": "https://inference-docs.cerebras.ai/api-reference/chat-completions",
"final_url": "https://inference-docs.cerebras.ai/api-reference/chat-completions",
"status_code": 200,
"title": "Chat Completions - Cerebras Inference",
"description": "Generate conversational responses using a structured message format with roles (system, user, assistant, developer, tool). Best for chatbots, assistants, and multi-turn conversations.",
"checked_at": "2026-10-24T17:44:44.8693014\u002B00:00",
"warnings": [
"\u6765\u6E90\u53EF\u8BFB\u53D6\u53EA\u8BC1\u660E\u672C\u6B21\u8D44\u6599\u8BFB\u53D6\uFF0C\u4E0D\u662F\u529F\u80FD\u3001\u8EAB\u4EFD\u6216\u957F\u671F\u53EF\u7528\u6027\u8BA4\u8BC1\u3002"
]
},
"created_at": "2026-10-24T17:44:48.5012668\u002B00:00",
"updated_at": "2026-10-24T17:49:02.2465278\u002B00:00",
"read_url": "https://navonavo.com/api/v1/capabilities/cap-67a2f1e31211aa65422c0b60?version=2",
"page_url": "https://navonavo.com/capabilities/cap-67a2f1e31211aa65422c0b60?version=2",
"verification_scope": "verification\u4E3A\u6295\u7A3F\u8005\u4E0A\u62A5\uFF1Bsource_check\u4EC5\u4E3A\u672C\u7AD9\u6765\u6E90\u8BFB\u53D6\uFF0C\u4E0D\u5347\u7EA7\u4E3A\u529F\u80FD\u9A8C\u8BC1\u6216\u5B98\u65B9\u8BA4\u8BC1\u3002",
"library_material": {
"organization": "organized",
"semantic_state": "classified",
"semantic_category": "\u4EBA\u5DE5\u667A\u80FD\uFF0FCerebras \u804A\u5929\u63A8\u7406\u4E0E\u5DE5\u5177\u8C03\u7528\u63A5\u53E3",
"evidence_grade": "documentation_only",
"attachments": [
{
"material_id": "material-b43916c920c546ba639b8cc24b82e169",
"source_url": "https://inference-docs.cerebras.ai/api-reference/chat-completions",
"content_sha256": "c0ee0b271dcdec5a1637c17b4c46cf639c8edf2c4696a11af5f17d69e6fa8f15",
"decoded_bytes": 751941,
"content_type": "text/html",
"captured_at": "2026-10-24T17:44:44.8693014\u002B00:00",
"text_representation": "decoded_text_utf8",
"text_utf8_sha256": "c0ee0b271dcdec5a1637c17b4c46cf639c8edf2c4696a11af5f17d69e6fa8f15",
"text_utf8_bytes": 751941,
"capture_body_reconstructed": true,
"read_path": "/api/v1/capabilities/cap-67a2f1e31211aa65422c0b60/materials/material-b43916c920c546ba639b8cc24b82e169?version=1"
}
]
},
"last_correction": {
"contribution_id": "contribution-64547e73c1a9461680e33019694fc9db",
"reason": "\u4F9D\u636E\u672C\u6279 Cerebras Chat Completions \u6B63\u6587\u53CA\u9644\u4EF6\u5185\u5D4C\u5B98\u65B9 bash \u793A\u4F8B\u6838\u5BF9\u5B8C\u6574 URL\u3001\u5FC5\u586B\u6D88\u606F\u5B57\u6BB5\u3001Bearer key \u548C\u8FD4\u56DE\u7ED3\u6784\uFF0C\u79FB\u9664\u672A\u6574\u7406\u6807\u7B7E\u3002\u53EA\u505A documentation_only\uFF0C\u8D39\u7528\u3001\u6570\u503C\u914D\u989D\u548C\u4FDD\u7559\u653F\u7B56\u4FDD\u6301\u672A\u77E5\u3002",
"source_urls": [
"https://inference-docs.cerebras.ai/api-reference/chat-completions"
],
"reported_by": "contributor"
},
"version_read_basis": "archived_revision",
"deep_read": {
"schema_version": "navonavo.capability-read.v1",
"card_id": "cap-67a2f1e31211aa65422c0b60",
"card_version": 2,
"selected_profile_id": null,
"state": "not_checked",
"review_state": null,
"read_url": "https://navonavo.com/api/v1/capabilities/cap-67a2f1e31211aa65422c0b60?version=2",
"page_url": "https://navonavo.com/capabilities/cap-67a2f1e31211aa65422c0b60?version=2",
"profiles": [],
"warnings": [
"\u8D44\u6599\u4E0E\u6765\u6E90\u58F0\u660E\u4E0D\u7B49\u4E8E\u672C\u7AD9\u5B9E\u6D4B\uFF1B\u672A\u5217\u51FA\u7684\u6761\u4EF6\u4ECD\u672A\u77E5\u3002",
"\u7ED3\u6784\u5316\u914D\u65B9\u53EA\u6309\u516C\u5F00\u58F0\u660E\u68C0\u67E5\u7EA6\u675F\u4E0E\u63A5\u53E3\uFF0C\u4E0D\u6267\u884C\u4EFB\u52A1\u6216\u6388\u4E88\u6743\u9650\uFF1B\u7F3A\u5931\u8BB0\u5F55\u4FDD\u6301\u672A\u77E5\uFF0C\u63A5\u53E3\u517C\u5BB9\u4E0D\u8BC1\u660E\u5B9E\u9645\u6210\u529F\u3002"
]
}
}